Using the known method of blind source separation BSS, a plurality of spatially distributed speakers who are talking simultaneously can be separated and emphasis can be placed on one of them in order to improve his/her audibility. BSS is disclosed in more detail in the publication EP 1 017 253 A2, whereby the useful acoustic source wanted by the hearing aid wearer is highlighted from a signal mixture. This or similar methods nevertheless require an adaptation time of several seconds, until the directions of the speakers can be reliably estimated. Because of unavoidable head movements which each time necessitate a renewed adaptation, these algorithms are thus currently only suitable for hearing aids to a limited extent.